The mammalian Y chromosome carries a factor that initiates male sexual development by directing the fetal gonads to form testes. Wachtel and his colleagues proposed that this testis-determining function of the Y is mediated by the male-specific cell-surface antigen H-Y, originally defined by skin grafting. This attractive hypothesis, which has been widely accepted, was based on the assumption that serological tests using antisera raised against male cells were recognizing H-Y antigen. Although disputed this assumption is supported by some recent studies. However, mice have been described which develop testes but lack the cell-surface H-Y antigen as defined by T-cell-mediated transplantation tests. Thus, although it remains possible that a serologically detected male-specific antigen is responsible for testis determination, it seems that H-Y, as originally defined, is not. We show here that H-Y negative male mice, in losing the genetic information that encodes H-Y, have also lost genetic information required for spermatogenesis. This result identifies a gene on the mouse Y, distinct from the testis-determining gene, which is necessary for spermatogenesis, and raises the intriguing possibility that the product of this 'spermatogenesis gene' is H-Y antigen.